BondNet: Noise-Robust Molecular Bond Perception from 3D Coordinates with Explicit Hydrogen Context
- 1School of Computer Science and Artificial Intelligence
- 2Shangqiu Normal University
Abstract
Recovering molecular graphs from atomic identities and 3D coordinates is a key step in coordinate-first molecular modelling and generation. We present Bond-Net, a two-stage framework in which a PaiNN model first predicts connectivity on a distance-derived candidate graph and a topology-aware network then classifies heavy–heavy bond types. Explicit hydrogens participate in message passing as local valence context but are excluded from bond-type supervision and scoring. On a fixed 26,940-molecule GEOM-DRUGS validation cohort, the near-clean model achieves pipeline macro-F1 = 0.9994 and correctly types every reference heavy–heavy bond in 99.4% of molecules. A historical ablation shows that removing hydrogen context reduces macro-F1 by 6.7 percentage points and the exact-type rate to 19.7%. With Gaussian coordinate-noise augmentation and candidate graphs rebuilt from the perturbed coordinates, BondNet retains pipeline macro-F1 = 0.9859 and a true-bond exact-type rate of 87.7% at σ = 0.1 Å; its stricter HH-graph exact match is 83.0% because extra predicted bonds are also penalized. BondNet’s pipeline macro-F1 exceeds RDKit DetermineBonds at every tested noise level on the common denominator. Residual strict-label disagreements are concentrated in resonance-ambiguous N–O and carboxyl motifs. An exact-molecule-leakage-controlled diversity study further shows that a curated multi-source set of 80,000 molecules transfers better than both a same-size random GEOM subset and a 3.4× larger single-source set. 1 Scientific Contribution. BondNet isolates geometric connectivity from topology-aware bond typing and quantifies how explicit-hydrogen context and noise augmentation affect robust graph recovery. The study also provides atom-pair-matched baselines, explicitly reports tool failures, and introduces an exact-molecule-leakage-controlled comparison of training-set diversity and scale.
Related
- Translation-rotation internal coordinates with splicing (TRICS): A new approach to molecular geometry interpolation
- Tensor-based Approximation of Molecular Kinetics: Generator Learning, Reaction Coordinates and Incremental Updating
- Ab Initio Molecular Dynamics with Explicit Solvation Reveals the Volmer–Heyrovsky Mechanism for Chlorine Evolution on Dimensionally Stable Anodes
- Potential Driven Spin-Crossover Synergized with Hydrogen-Bond Engineering in a Molecular Catalyst for Selective H2O2 Electrosynthesis.
- Hydrogen diffusion on metals: Full-Dimensional Ring Polymer Molecular Dynamics Rate Calculations
BibTeX
@misc{zhang2026bondnet,
title = {{BondNet: Noise-Robust Molecular Bond Perception from 3D Coordinates with Explicit Hydrogen Context}},
author = {Xiaochen Zhang and Hui Zeng and Xiao-Chen Zhang},
howpublished = {ChemRxiv},
year = {2026},
month = sep,
doi = {10.26434/chemrxiv.15008897/v1},
url = {https://chemrxiv.org/doi/full/10.26434/chemrxiv.15008897/v1},
note = {CC BY},
}
RIS
TY - UNPB
TI - BondNet: Noise-Robust Molecular Bond Perception from 3D Coordinates with Explicit Hydrogen Context
AU - Xiaochen Zhang
AU - Hui Zeng
AU - Xiao-Chen Zhang
PY - 2026
DA - 2026/09/16
PB - ChemRxiv
DO - 10.26434/chemrxiv.15008897/v1
UR - https://chemrxiv.org/doi/full/10.26434/chemrxiv.15008897/v1
AB - Recovering molecular graphs from atomic identities and 3D coordinates is a key step in coordinate-first molecular modelling and generation. We present Bond-Net, a two-stage framework in which a PaiNN model first predicts connectivity on a distance-derived candidate graph and a topology-aware network then classifies heavy–heavy bond types. Explicit hydrogens participate in message passing as local valence context but are excluded from bond-type supervision and scoring. On a fixed 26,940-molecule GEOM-DRUGS validation cohort, the near-clean model achieves pipeline macro-F1 = 0.9994 and correctly types every reference heavy–heavy bond in 99.4% of molecules. A historical ablation shows that removing hydrogen context reduces macro-F1 by 6.7 percentage points and the exact-type rate to 19.7%. With Gaussian coordinate-noise augmentation and candidate graphs rebuilt from the perturbed coordinates, BondNet retains pipeline macro-F1 = 0.9859 and a true-bond exact-type rate of 87.7% at σ = 0.1 Å; its stricter HH-graph exact match is 83.0% because extra predicted bonds are also penalized. BondNet’s pipeline macro-F1 exceeds RDKit DetermineBonds at every tested noise level on the common denominator. Residual strict-label disagreements are concentrated in resonance-ambiguous N–O and carboxyl motifs. An exact-molecule-leakage-controlled diversity study further shows that a curated multi-source set of 80,000 molecules transfers better than both a same-size random GEOM subset and a 3.4× larger single-source set. 1 Scientific Contribution. BondNet isolates geometric connectivity from topology-aware bond typing and quantifies how explicit-hydrogen context and noise augmentation affect robust graph recovery. The study also provides atom-pair-matched baselines, explicitly reports tool failures, and introduces an exact-molecule-leakage-controlled comparison of training-set diversity and scale.
ER -